Add Solidity language support

Cargo.toml

@@ -44,6 +44,7 @@ tree-sitter-c-sharp = { version = "0.23.1" }
 tree-sitter-html = "0.23.2"
 tree-sitter-md = "0.3.2"
 tree-sitter-yaml = "0.6.1"
+tree-sitter-solidity = "=1.2.10"
 ast-grep-core = "0.36.1"
 ast-grep-language = "0.36.1"
 clap = { version = "4.3", features = ["derive"] }

README.md

@@ -143,7 +143,7 @@ npx -y @probelabs/probe@latest agent "Refactor the login function" --allow-edit
 - **Smart Ranking**: BM25, TF-IDF, and hybrid algorithms with optional BERT reranking
 - **Token-Aware**: `--max-tokens` budget, session-based dedup to avoid repeating context
 - **Built-in Agent**: Multi-provider (Anthropic, OpenAI, Google, Bedrock) with retry, fallback, and context compaction
-- **Multi-Language**: Rust, Python, JavaScript, TypeScript, Go, C/C++, Java, Ruby, PHP, Swift, C#, and more
+- **Multi-Language**: Rust, Python, JavaScript, TypeScript, Go, C/C++, Java, Ruby, PHP, Swift, Solidity, C#, and more
 
 ---
 
@@ -468,6 +468,7 @@ cargo install --path .
 | Ruby | `.rb` |
 | PHP | `.php` |
 | Swift | `.swift` |
+| Solidity | `.sol` |
 | C# | `.cs` |
 | Markdown | `.md` |
 

docs/probe-cli/extraction-reference.md

@@ -177,6 +177,7 @@ Probe supports extraction for many other languages including:
 - **Ruby**: Methods, classes, modules, blocks
 - **PHP**: Functions, classes, namespaces, attributes
 - **Swift**: Functions, classes, structs, protocols, extensions
+- **Solidity**: Contracts, interfaces, libraries, functions, modifiers, events, errors
 - **C#**: Methods, classes, interfaces, namespaces, attributes
 - **Markdown**: Sections, code blocks, lists, tables, frontmatter
 
@@ -433,4 +434,4 @@ probe extract src/api.js:27 --context 10
 - **INTELLIGENT FALLBACKS**: Gracefully handles cases where AST parsing isn't possible
 
 For more information on how Probe works internally, see [How Probe Works](how-it-works.md).
-For details on search capabilities, see [Search Functionality](search-functionality.md).
+For details on search capabilities, see [Search Functionality](search-functionality.md).

docs/probe-cli/query.md

@@ -138,6 +138,7 @@ Required for accurate parsing:
 | Ruby | `ruby`, `rb` | .rb |
 | PHP | `php` | .php |
 | Swift | `swift` | .swift |
+| Solidity | `solidity`, `sol` | .sol |
 | C# | `csharp`, `cs` | .cs |
 
 ```bash

docs/probe-cli/search.md

@@ -208,7 +208,7 @@ probe search "api" ./ --reranker ms-marco-tinybert --question "How is the REST A
 | `-l`, `--language` | String | auto | Limit to programming language |
 
 **Supported Languages:**
-`rust`, `javascript`, `typescript`, `python`, `go`, `c`, `cpp`, `java`, `ruby`, `php`, `swift`, `csharp`, `yaml`, `html`, `markdown`
+`rust`, `javascript`, `typescript`, `python`, `go`, `c`, `cpp`, `java`, `ruby`, `php`, `swift`, `solidity`, `csharp`, `yaml`, `html`, `markdown`
 
 ```bash
 # Search only Python files

docs/reference/adding-languages.md

@@ -95,7 +95,7 @@ Add to language completion lists (2 locations):
 ### File 10: `npm/src/mcp/index.ts`
 Add to supported languages description:
 ```rust
-'Supported languages: rust, javascript, typescript, python, go, c, cpp, java, ruby, php, swift, csharp, yourlang.',
+'Supported languages: rust, javascript, typescript, python, go, c, cpp, java, ruby, php, swift, solidity, csharp, yourlang.',
 ```
 
 ## Step 4: Documentation

docs/reference/architecture.md

@@ -104,7 +104,7 @@ pub trait LanguageImpl {
 }
 ```
 
-**Supported:** Rust, JavaScript, TypeScript, Python, Go, C, C++, Java, Ruby, PHP, Swift, C#, HTML, Markdown, YAML
+**Supported:** Rust, JavaScript, TypeScript, Python, Go, C, C++, Java, Ruby, PHP, Swift, Solidity, C#, HTML, Markdown, YAML
 
 ### Performance Optimizations
 
@@ -418,4 +418,3 @@ agent.events.on('toolCall', (event) => {
 - [API Reference](../probe-agent/sdk/api-reference.md) - SDK API
 - [MCP Protocol](../probe-agent/protocols/mcp.md) - MCP integration
 - [Performance](../probe-cli/performance.md) - Optimization guide
-

docs/reference/faq.md

@@ -62,7 +62,7 @@ Yes, Probe is open source under the Apache 2.0 license.
 
 ### What languages does Probe support?
 
-Probe supports 15+ languages including Rust, JavaScript, TypeScript, Python, Go, C, C++, Java, Ruby, PHP, Swift, C#, HTML, Markdown, and YAML. See [Supported Languages](../supported-languages.md).
+Probe supports 15+ languages including Rust, JavaScript, TypeScript, Python, Go, C, C++, Java, Ruby, PHP, Swift, Solidity, C#, HTML, Markdown, and YAML. See [Supported Languages](../supported-languages.md).
 
 ---
 

docs/reference/language-support.md

@@ -107,6 +107,7 @@ Probe currently supports a wide range of programming languages, including:
 - Ruby
 - PHP
 - Swift
+- Solidity
 - C#
 - Markdown
 
@@ -165,4 +166,4 @@ Tree-sitter provides several advantages for Probe:
 - [Supported Languages](/supported-languages): Detailed information about each supported language
 - [Adding New Languages](/adding-languages): Guide for adding support for new languages
 - [Search Functionality](/search-functionality): How to use Probe's search features
-- [Code Extraction](/code-extraction): Details on how Probe extracts code blocks
+- [Code Extraction](/code-extraction): Details on how Probe extracts code blocks

docs/reference/supported-languages.md

@@ -18,6 +18,7 @@ Probe provides language-aware code search and extraction for a wide range of pro
 | Ruby | `.rb` | ✅ | ✅ |
 | PHP | `.php` | ✅ | ✅ |
 | Swift | `.swift` | ✅ | ✅ |
+| Solidity | `.sol` | ✅ | ✅ |
 | C# | `.cs` | ✅ | ✅ |
 | Markdown | `.md`, `.markdown` | ✅ | ✅ |
 | YAML | `.yaml`, `.yml` | ✅ | ✅ |
@@ -128,6 +129,14 @@ Go also implements special handling for nested struct types.
 - **Attribute Handling**: Properly handles C# attributes
 - **Test Detection**: Identifies test classes and methods using NUnit, MSTest, or xUnit conventions
 
+### Solidity
+
+- **Contract Extraction**: Extracts contracts, interfaces, and libraries
+- **Function Extraction**: Extracts functions, constructors, fallback/receive functions, and modifiers
+- **Type Extraction**: Extracts structs, enums, user-defined value types, events, and custom errors
+- **Comment Handling**: Associates NatSpec and regular comments with Solidity code blocks
+- **Test Detection**: Identifies Foundry-style `.t.sol` files, `*Test.sol` contracts, `setUp`, `test*`, and `invariant_*` functions
+
 ### Markdown
 
 - **Section Extraction**: Extracts complete sections based on headings
@@ -212,4 +221,4 @@ func ($RECEIVER $TYPE) $NAME($$$PARAMS) $$$BODY
 
 ## Adding Support for New Languages
 
-If you're interested in adding support for a language that's not currently supported, see the [Adding New Languages](/adding-languages) page for a detailed guide.
+If you're interested in adding support for a language that's not currently supported, see the [Adding New Languages](/adding-languages) page for a detailed guide.

lsp-daemon/Cargo.toml

@@ -81,6 +81,7 @@ tree-sitter-c-sharp = "0.23.1"
 tree-sitter-ruby = "0.23.1"
 tree-sitter-swift = "0.7.0"
 tree-sitter-php = "0.23.11"
+tree-sitter-solidity = "=1.2.10"
 
 [target.'cfg(unix)'.dependencies]
 libc = "0.2"

lsp-daemon/src/analyzer/tree_sitter_analyzer.rs

@@ -31,6 +31,7 @@
         "php" => Some("php"),
         "swift" => Some("swift"),
         "cs" => Some("csharp"),
+        "sol" => Some("solidity"),
         _ => None,
     }
 }
@@ -56,7 +57,7 @@
             extension_to_language_name(language_or_extension).unwrap_or(language_or_extension);
 
         let language_parsers = self
             .parsers
             .entry(language_name.to_string())
             .or_insert_with(Vec::new);
 
@@ -98,6 +99,7 @@
             "java" => Some(tree_sitter_java::LANGUAGE),
             "c" => Some(tree_sitter_c::LANGUAGE),
             "cpp" | "c++" | "cxx" => Some(tree_sitter_cpp::LANGUAGE),
+            "solidity" | "sol" => Some(tree_sitter_solidity::LANGUAGE),
             _ => None,
         };
 

lsp-daemon/src/daemon.rs

@@ -7214,9 +7214,15 @@
             "cpp" | "cc" | "cxx" | "hpp" => {
                 parser
                     .set_language(&tree_sitter_cpp::LANGUAGE.into())
                     .ok()?;
                 Some(())
             }
+            "sol" => {
+                parser
+                    .set_language(&tree_sitter_solidity::LANGUAGE.into())
+                    .ok()?;
+                Some(())
+            }
             _ => None,
         }?;
 

lsp-daemon/src/fqn.rs

@@ -41,9 +41,10 @@
         "js" | "jsx" => Some(tree_sitter_javascript::LANGUAGE),
         "ts" | "tsx" => Some(tree_sitter_typescript::LANGUAGE_TYPESCRIPT),
         "java" => Some(tree_sitter_java::LANGUAGE),
         "go" => Some(tree_sitter_go::LANGUAGE),
         "c" => Some(tree_sitter_c::LANGUAGE),
         "cpp" | "cc" | "cxx" => Some(tree_sitter_cpp::LANGUAGE),
+        "sol" => Some(tree_sitter_solidity::LANGUAGE),
         _ => None,
     };
 
@@ -357,10 +358,10 @@
 }
 
 /// Get language-specific separator for FQN components
 fn get_language_separator(extension: &str) -> &str {
     match extension {
         "rs" | "cpp" | "cc" | "cxx" | "hpp" | "hxx" | "rb" => "::",
-        "py" | "js" | "ts" | "jsx" | "tsx" | "java" | "go" | "cs" => ".",
+        "py" | "js" | "ts" | "jsx" | "tsx" | "java" | "go" | "cs" | "sol" => ".",
         "php" => "\\",
         _ => "::", // Default to Rust-style for unknown languages
     }
@@ -381,6 +382,13 @@
         "java" | "cs" => kind == "method_declaration",
         "go" => kind == "function_declaration",
         "cpp" | "cc" | "cxx" => matches!(kind, "function_definition" | "method_declaration"),
+        "sol" => matches!(
+            kind,
+            "function_definition"
+                | "constructor_definition"
+                | "modifier_definition"
+                | "fallback_receive_definition"
+        ),
         _ => kind.contains("function") || kind.contains("method"),
     }
 }
@@ -403,6 +411,14 @@
             kind,
             "class_specifier" | "struct_specifier" | "namespace_definition"
         ),
+        "sol" => matches!(
+            kind,
+            "contract_declaration"
+                | "interface_declaration"
+                | "library_declaration"
+                | "struct_declaration"
+                | "enum_declaration"
+        ),
         _ => {
             // Fallback for unknown languages: try to detect common node types
             kind.contains("class") || kind.contains("struct") || kind.contains("namespace")

lsp-daemon/src/indexing/ast_extractor.rs

@@ -697,6 +697,9 @@
             crate::language_detector::Language::Java => Ok(tree_sitter_java::LANGUAGE.into()),
             crate::language_detector::Language::C => Ok(tree_sitter_c::LANGUAGE.into()),
             crate::language_detector::Language::Cpp => Ok(tree_sitter_cpp::LANGUAGE.into()),
+            crate::language_detector::Language::Solidity => {
+                Ok(tree_sitter_solidity::LANGUAGE.into())
+            }
             _ => Err(anyhow::anyhow!("Unsupported language: {:?}", language)),
         }
     }
@@ -831,9 +834,23 @@
                 "method_declaration" | "constructor_declaration" => (SymbolKind::Function, true),
                 "class_declaration" => (SymbolKind::Class, true),
                 "interface_declaration" => (SymbolKind::Interface, true),
                 "field_declaration" => (SymbolKind::Variable, true),
                 _ => (SymbolKind::Function, false),
             },
+            crate::language_detector::Language::Solidity => match node_kind {
+                "function_definition" | "fallback_receive_definition" => {
+                    (SymbolKind::Function, true)
+                }
+                "constructor_definition" | "modifier_definition" => (SymbolKind::Method, true),
+                "contract_declaration" | "library_declaration" => (SymbolKind::Class, true),
+                "interface_declaration" => (SymbolKind::Interface, true),
+                "struct_declaration" => (SymbolKind::Struct, true),
+                "enum_declaration" => (SymbolKind::Enum, true),
+                "event_definition" | "error_declaration" => (SymbolKind::Type, true),
+                "state_variable_declaration" => (SymbolKind::Variable, true),
+                "user_defined_type_definition" => (SymbolKind::Type, true),
+                _ => (SymbolKind::Function, false),
+            },
             _ => {
                 // For other languages, try some common patterns
                 match node_kind {
@@ -930,10 +947,20 @@
                     }
                 }
                 _ => continue,
             }
         }
 
-        None
+        match node.kind() {
+            "constructor_definition" => Some("constructor".to_string()),
+            "fallback_receive_definition" => node.utf8_text(content).ok().map(|text| {
+                if text.trim_start().starts_with("receive") {
+                    "receive".to_string()
+                } else {
+                    "fallback".to_string()
+                }
+            }),
+            _ => None,
+        }
     }
 }
 

lsp-daemon/src/indexing/config.rs

@@ -1401,6 +1401,7 @@
         Language::Java,
         Language::C,
         Language::Cpp,
+        Language::Solidity,
     ] {
         let lang_str = format!("{language:?}").to_uppercase();
         let mut config = LanguageIndexConfig::default();
@@ -1498,6 +1499,11 @@
                         features.set_language_feature("extract_namespaces".to_string(), true);
                         features.set_language_feature("extract_classes".to_string(), true);
                     }
+                    Language::Solidity => {
+                        features.set_language_feature("extract_contracts".to_string(), true);
+                        features.set_language_feature("extract_events".to_string(), true);
+                        features.set_language_feature("extract_modifiers".to_string(), true);
+                    }
                     _ => {}
                 }
 
@@ -1518,7 +1524,7 @@
 fn parse_bool_env(value: &str, var_name: &str) -> Result<bool> {
     match value.to_lowercase().as_str() {
         "true" | "1" | "yes" | "on" | "enabled" => Ok(true),
         "false" | "0" | "no" | "off" | "disabled" => Ok(false),
         _ => Err(anyhow!("Invalid boolean value for {}: {} (use true/false, 1/0, yes/no, on/off, enabled/disabled)", var_name, value)),
     }
 }
@@ -1540,6 +1546,7 @@
             "hpp".to_string(),
             "hxx".to_string(),
         ],
+        Language::Solidity => vec!["sol".to_string()],
         _ => vec![],
     }
 }
@@ -1557,6 +1564,7 @@
             "java" => Ok(Language::Java),
             "c" => Ok(Language::C),
             "cpp" | "c++" => Ok(Language::Cpp),
+            "solidity" | "sol" => Ok(Language::Solidity),
             _ => Err(anyhow!("Unknown language: {}", s)),
         }
     }

lsp-daemon/src/indexing/file_detector.rs

@@ -122,6 +122,7 @@ impl Default for DetectionConfig {
             "rb",
             "php",
             "swift",
+            "sol",
             "cs",
             "kt",
             "scala",
@@ -384,9 +385,9 @@ impl FileChangeDetector {
             // Check if this extension is supported based on our known languages
             let supported_languages = [
                 "rs", "js", "jsx", "ts", "tsx", "py", "go", "c", "h", "cpp", "cc", "cxx", "hpp",
-                "hxx", "java", "rb", "php", "swift", "cs", "kt", "scala", "clj", "ex", "exs",
-                "erl", "hrl", "hs", "lhs", "ml", "mli", "fs", "fsx", "fsi", "dart", "jl", "r", "R",
-                "m", "mm", "pl", "pm", "sh", "bash", "zsh", "fish", "lua", "vim", "sql",
+                "hxx", "java", "rb", "php", "swift", "sol", "cs", "kt", "scala", "clj", "ex",
+                "exs", "erl", "hrl", "hs", "lhs", "ml", "mli", "fs", "fsx", "fsi", "dart", "jl",
+                "r", "R", "m", "mm", "pl", "pm", "sh", "bash", "zsh", "fish", "lua", "vim", "sql",
             ];
 
             if supported_languages.contains(&extension) {

lsp-daemon/src/language_detector.rs

@@ -20,6 +20,7 @@ pub enum Language {
     Ruby,
     Php,
     Swift,
+    Solidity,
     Kotlin,
     Scala,
     Haskell,
@@ -45,6 +46,7 @@ impl Language {
             Language::Ruby => "ruby",
             Language::Php => "php",
             Language::Swift => "swift",
+            Language::Solidity => "solidity",
             Language::Kotlin => "kotlin",
             Language::Scala => "scala",
             Language::Haskell => "haskell",
@@ -72,6 +74,7 @@ impl Language {
             "ruby" => Some(Language::Ruby),
             "php" => Some(Language::Php),
             "swift" => Some(Language::Swift),
+            "solidity" | "sol" => Some(Language::Solidity),
             "kotlin" => Some(Language::Kotlin),
             "scala" => Some(Language::Scala),
             "haskell" => Some(Language::Haskell),
@@ -146,6 +149,9 @@ impl LanguageDetector {
         // Swift
         extension_map.insert("swift".to_string(), Language::Swift);
 
+        // Solidity
+        extension_map.insert("sol".to_string(), Language::Solidity);
+
         // Kotlin
         extension_map.insert("kt".to_string(), Language::Kotlin);
         extension_map.insert("kts".to_string(), Language::Kotlin);